Hello kernel-discuss,
Few things I have in my queue:
1. irq_base transition. Some time ago asic3_base was changed to new
virtual (multiplexed) interrupt allocation scheme, where IRQ base for
them is passed via platform_data, instead of being allocated by an
adhoc function, not part of mainline. This was highly successful, and
I'm looking into converting all other drivers to it.
The question here is if we should use struct resource for passing
this irq_base? (Chip's own IRQ is already passed via it, and of course
the whole idea of struct resource is that it should be used whenever
possible). Thoughts?
2. Location of SOC drivers include files. We had include/linux/soc/
for this, before drivers/soc/ was moved to drivers/mfd/. Now
include/linux/mfd/ was opened by htc-egpio.h. Obviously, they all
should be kept in one place, so I'm looking into moving old ones to
<linux/mfd/>. Caveat is that it's still unclear if that would be
opened in mainline (it seems they just have fun watching how
include/linux/ approaches 1K files).
3. Codestyle for new code. Well, we already on stage of cleanup of
code for mainline conformity, so it would be nice if we took older
lessons and write code in such a way that we don't need to
rewrite/reformat later again (pretty boring task). In this regard,
htc-egpio.* has inconsistent symbol naming scheme comparing with the
mainline kernel (and doesn't even use that scheme consistently).
Kevin, could you look into this?
Thanks,
-- Best regards, Paul mailto:pmiscml_at_gmail.comReceived on Mon Jul 02 2007 - 10:31:42 EDT
This archive was generated by hypermail 2.2.0 : Mon Jul 02 2007 - 10:31:59 EDT